Do people not wire wrap anymore?

[Beamin]

I know when I was a kid back in the 90's that's how I made everything. Still today I do that. Although I see almost no one doing that on you tube. Also just buying the little tool cost 30.00$ just for one size wire! I remember the tools used to be throw away cheap too not 60$ for the two sizes you need. Is it because now everyone has laser printers and at home PCBs?

[gnif]

I never learnt to use wire wrap, but did see examples of it in my dad's electronics junk box when I was a child. I went straight for strip board and kynar wire for prototyping, all soldered. Technically not wire wrap but using the wire

[georges80]

1) Most modern components are surface mount.
2) Modern components often run at much higher clock speeds thnt the 'good old days'.
3) Cost of high quality PCBs is considerably cheaper than the 'good old days' and the design/layout tools are cheap/free. Turn around time is also just days or a week at most.

There are other reasons but the above pretty well makes wirewrapping obsolete. I'm glad to have left that way behind me. I wirewrapped many a z80 based computer way back when. Also a few 16bit/32bit machines. Getting a machine running reliably at 10MHz was challenging enough.

For the cost of a wirewrap tool + wirewrap sockets + decent wirewrap board you could do a PCB and get it delivered for less money and a better result. I'd rather do the schematic/layout than sit for hours going zip zip zip with an electric wirewrap gun.... With a PCB I can build 2 units or 3 units or .... The only use I have for wirewrap wire is for reworking/tweaking a design.

cheers,
george.

[NiHaoMike]

And instead of building logic circuits out of individual 74xx chips, you can write some Verilog or VHDL and load it onto a FPGA.

[jmelson]

I still make test fixtures and adapters with wire wrap.  But, for anything with more than a couple ICs worth of logic, I will repurpose one of my CPLD or FPGA boards to do all the logic, but might use wire-wrap to connect it to the outside world.

There's just no sense to wire up a bunch of 74xx chips today.

Jon

[ajb]

I've wire-wrapped a couple of small things, but obviously you can't wire wrap SMT parts without a bunch of adapter boards--plus wirewrap sockets are expensive these days!  So for perfboard stuff I much prefer the ELM method.  I usually use a DIP-form factor carrier PCB for the MCU--there are plenty of cheapo ebay options for STM32 parts, and I've designed a few myself for others.

But really, anything of any complexity or that I might ever want to do more than one of gets a custom PCB.

[Brumby]

I moved from the spring clips of the Philips EE20 kit to Veroboard and soldering iron.

I had seen wire wrap, but wasn't too sure about it.  That was until I read an article that described the process and the reasoning behind it which gave clarification to the reliability aspect.

To this day I have never done any wire wrapping (well, of the kind discussed here  ), I don't have - and never have had - any such wire or a tool.  I will admit to having some wire-wrap sockets, though.  (They came in a grab bag.)

[NiHaoMike]

There's just no sense to wire up a bunch of 74xx chips today.

I'm doing just that to make a controller for a bidirectional DC/DC converter. (Actually using 4000 series because that part of the logic runs on 12V.) A FPGA or CPLD would not be able to meet the very low standby/pass through power goal of 1mA or less, plus it would require level translation.

[thermistor-guy]

I'm using wire-wrap now to build some custom test jigs, for some low-frequency calibration work. About 500 connections in total. Since this is a prototype, the connections must be reliable but changeable - wire-wrap meets the requirements. The connectors are fastened directly to steel plate (for various reasons). There is no PCB in the jig.

But back in the day, I saw high-speed ECL graphics cards that were wire-wrapped (two-level, not the usual three-level). The cards were used in industrial minicomputers.  If you know what you are doing, you can use wire-wrap it for high speed (ground plane construction, Schottky diode terminations on long clock lines, etc.).

[floobydust]

It was expensive (IC sockets, WW tool) and then WW a capacitor or resistor and you might as well just use solder anyway.
A PLCC wire-wrap socket was very hard to find. 28-30AWG seems to crap out at around 20MHz for CPU bus I did.
Wire-wrap was popular amongst students years ago, for project courses.

[IanMacdonald]

Wire wrapping was another of those products which was 'hyped' without being properly evaluated. Its key problem is that if a connection needs changing then any other wires on top of that connection have to be removed first, and replaced after. Which can soon develop into a domino effect since the wires removed are now too short to be reused, so any wires on top of the other ends also need to be replaced, and so on.

[daveshah]

I'm doing just that to make a controller for a bidirectional DC/DC converter. (Actually using 4000 series because that part of the logic runs on 12V.) A FPGA or CPLD would not be able to meet the very low standby/pass through power goal of 1mA or less, plus it would require level translation.

The level translation issue I accept, but for future reference something like a MachXO2 or an iCE40 have static currents way below 1mA (<100µA), so FPGAs aren't always power hungry.

[NiHaoMike]

I'm doing just that to make a controller for a bidirectional DC/DC converter. (Actually using 4000 series because that part of the logic runs on 12V.) A FPGA or CPLD would not be able to meet the very low standby/pass through power goal of 1mA or less, plus it would require level translation.

The level translation issue I accept, but for future reference something like a MachXO2 or an iCE40 have static currents way below 1mA (<100µA), so FPGAs aren't always power hungry.

Plus the quiescent current of the regulator and level translation circuits. Not to mention that the part of the logic I'm implementing in 4000 series (just a few gates plus two flip flops) is very simple and a FPGA would be overkill. The function is very simple - just pass through the main drive signal (from the PIC via a level shifter) for the mode of operation while the synchronous rectification signal is gated off by the current crossing (near) zero until the next cycle. And implement overcurrent protection for the MOSFETs that works by gating off the drive signals and signaling to the PIC that an overcurrent trip has occurred.

[tggzzz]

Wire wrapping was another of those products which was 'hyped' without being properly evaluated. Its key problem is that if a connection needs changing then any other wires on top of that connection have to be removed first, and replaced after. Which can soon develop into a domino effect since the wires removed are now too short to be reused, so any wires on top of the other ends also need to be replaced, and so on.

No, it wasn't hyped. Wirewrapping was designed to be a high reliability interconnection technique that could be modified on infrequent occasions. It succeeded in those objectives.

It had limitations in the prototyping environment, particularly for those people that connected first and thouht afterwards. With thought and care it wasn't too bad for prototyping, and certainly more accessible than PCBs in the 70s and 80s.

The real killer was the pins act as stubs, which is very problematic for high speed logic.

[tggzzz]

Wirewrap is a current production technique, e.g.


Wirewrap has been superceded for prototyping for many solid reasons.

[bob225]

Wire wrap still has its uses - I have just repaired a control board what was in 5 pieces, as I could not get a replacement pcb for love nor money

once in a blue moon I will bind cables together with it

[joeqsmith]

I've used WW many times over the years because of the high costs of PCBs.  In the 80's I had a slit wrap gun that used a Teflon wire.     When I decided to design and build a better transient generator, I thought it would be fun to revisit the past and use WW with an old Motorola controller.   

Several WW projects I built are still operational today.

[rstofer]

Wire wrapping was another of those products which was 'hyped' without being properly evaluated. Its key problem is that if a connection needs changing then any other wires on top of that connection have to be removed first, and replaced after. Which can soon develop into a domino effect since the wires removed are now too short to be reused, so any wires on top of the other ends also need to be replaced, and so on.

But, of course, we never wrapped it in a daisy chain.  We always wired the bottom level first (pin 1 to pin 2 and pin 3 to pin 4) and then the top level (pin 2 to pin 3).  Still, that doesn't make rework a lot of fun.

I go back a long way with computers ('70) and it was very common for major components (like line printers and card readers) to be wire-wrapped.  Among other things, the logic could be a lot denser with wire-wrap than with a PCB when those were made from taped layouts and photography

https://www.eetimes.com/author.asp?section_id=14&doc_id=1285442
http://www.computerhistory.org/projects/ibm_1620/ibm1620/PAGES/photos_miscellaneous.html

Even if the individual cards were printed, the backplane would still be wire-wrapped.

There was a time before CAD and things worked pretty well despite the lack.

I have built projects as large as 100 ICs with wire-wrap.  I don't want to do it again but I certainly have the Gardner-Denver Cut-Strip-Wrap gun for doing it.

I can wire-wrap a 32 bit register with 32 input wires, 32 output wires, a clock and reset signal to  8 4-bit register chips OR I can write one line of code to create the register and a wee bit more typing to 'wire' it up.

[jmelson]

Wire wrapping was another of those products which was 'hyped' without being properly evaluated. Its key problem is that if a connection needs changing then any other wires on top of that connection have to be removed first, and replaced after. Which can soon develop into a domino effect since the wires removed are now too short to be reused, so any wires on top of the other ends also need to be replaced, and so on.

IBM 360 and 370 computers had interconnect on the "boards" that held the circuit cards.  An additional layer of interconnect plus field changes and option jumpering was done in wire wrap.

Plug-compatible mainframes made by at least one company (National Advanced Systems) were all built out of massive wire-wrap panels covered with DIP chips.  So, the entire CPU was wire-wrapped.  Several minicomputers were also made with wire-wrap.

The phone company developed wire wrap in the 1950's (I think) for making connections between phone switches and the junctions to the subscriber wiring plant.

So, wire-wrap was NOT "hyped" it was a VERY mainstream manufacturing technology.

Jon

[jmelson]

Oh, some links !

http://pico-systems.com/stories/1982.html

http://pico-systems.com/stories/1982b.html

There are some larger projects that I wire-wrapped! (But, that was quite a while ago!)

Jon

[james_s]

Wire wrap is still used for building test fixtures, I visited a company not long ago that was using it. I don't think it gets used much for hobby prototyping since there are other easier options around but I'd still like to wire wrap a single board 8 bit computer sometime just for fun.

[georges80]

... and then there was multi-wire. Seems to still exist.

http://www.hitachi-chemical.com/products_pwb_05.htm

cheers,
george.

[Neilm]

Some test fixtures are still designed to use wire wrapping. I know at work the ICT units are wire wrapped.

[james_s]

Wasn't it widely used in telephone exchanges too? Those are pretty much a thing of the past but there were a whole lot of them up into the 1990s.

[GerryBags]

I've never used WW, but it seems as though it would be much easier to check lots of points with a 'scope than it is with a bread-boarded circuit. Once there are a few jumper wires dotted around the place it can hard to find somewhere to clip on a probe. Admittedly, this is without the experience necessary to know the best way to arrange common combos of ICs to optimize interconnecting them.

OT -  I see Dave's or Shariah's breadboards when they demo a circuit (such a useful thing) and they're so neat and tidy. I know they obviously spend extra time on that aspect as they know it's going on YT, but even so, I don't think there IS enough time for me to get a prototype circuit to look that smart.

I really regret the 40+ years of my life spent thinking that electronics was more akin to black magic than science, and that the whole subject was "beyond me". I feel like I started a little late.    Although, I know realise that the main problem was getting explanations of basic electrical concepts from people who didn't have a scooby-doo, though they professed to be experts. Inconsistencies are bound to arise. Until coming across people like Alan (W2aew), Shariah and Dave (and the NEETS course) it seemed that most of the people who do know what they're talking about didn't start at a basic enough level for me.

[thermistor-guy]

Wire wrapping was another of those products which was 'hyped' without being properly evaluated. Its key problem is that if a connection needs changing then any other wires on top of that connection have to be removed first, and replaced after. Which can soon develop into a domino effect since the wires removed are now too short to be reused, so any wires on top of the other ends also need to be replaced, and so on.

I disagree with all of this.

NASA has workmanship standards for wire-wrap:
https://workmanship.nasa.gov/lib/insp/2%20books/links/sections/301_discrete%20wiring.html

When wire-wrapping, you wrap wires between the same level on the posts, wherever possible. This limits any domino effect to a few wires at most. A second refinement is to wrap the longer connections first, then use the shorter connections (that run at a different angle), to "pin down" the longer wires close to the PCB or substrate surface.

With a little experience, it becomes second nature to do it this way. Done with care, it looks professional and performs well.

[joeqsmith]

Oh, some links !

http://pico-systems.com/stories/1982.html

http://pico-systems.com/stories/1982b.html

There are some larger products that I wire-wrapped! (But, that was quite a while ago!)

Jon

Very nice work! 

This was one of the last things I built using WW.   This box has a 6811 with some DRAM, FPGA and Ethernet controller.   It emulates an old Centronics printer port and sends the data to an printer server over Ethernet.         

https://youtu.be/l0GiC-dw9W8?list=PLZSS2ajxhiQCKHI3sA_DMpjdnIt9NiOb6&t=124

[joeqsmith]

Back in my analog scope days, I made this digitizer on WW.   

[Cubdriver]

Back in the mid-late 80s when I worked for National Semiconductor, we had a nitride coater made by Coyote Systems that had wire-wrapped circuit boards.  It was scary to look in the card cage, but it never gave us any trouble in that area.  (The vacuum and RF sections, on the other hand....)

HP used wire wrapping for the backplane of the 5360A Computing Counter in the early 70s, too.  (All discreet logic chips, around 300 of them all told.  They're all on small circuit boards, but part of the backplane is shown below):



Properly done, like crimping, it's a very reliable system - the sharp edges of the posts dig into the wire surface, foming a gas tight connection that makes for excellent contact.

-Pat

[tggzzz]

Properly done, like crimping, it's a very reliable system - the sharp edges of the posts dig into the wire surface, foming a gas tight connection that makes for excellent contact.

Just so, except that it is 28 gas-tight connections (assuming the coil has 7 turns).

[GerryBags]

Are the 30AWG wires comparable to PCB traces in terms of stay inductance/capacitance?

I notice Jim Williams included a pic of a large wire wrapped backplane in AN47 on high speed prototyping.

[tggzzz]

Are the 30AWG wires comparable to PCB traces in terms of stay inductance/capacitance?

There is more variability due to the varying distances between wires (i.e. crosstalk) and any groundplane.

The pins are 20mm long which means signals will ring with about 0.5ns period; not good with modern jellybean logic!
https://www.edn.com/electronics-blogs/all-aboard-/4438918/What-is-the-ringing-period-on-an-unterminated-line--Rule-of-Thumb--26

[woodchips]

For its era, WW was a perfectly good interconnect method. I used to WW telephone exchanges, millions of connections! Worked reliably, WW is more reliable than soldered joints. As has been mentioned, some simple rules of thumb, wraps at the same level, don't make the interconnect wires too tight, don't take all bus wires the same route etc and worked fine at 80's processor speeds, even video if careful. It was easy enough to unwrap a connection and then re-wrap it, could do it at least once.

But then Vero Speedwire arrived in the early 80's, which is what I have used ever since.

[tggzzz]

But then Vero Speedwire arrived in the early 80's, which is what I have used ever since.

For some reason I never got on with that.

I did like the IDC version of wirewrap. Imagine a 4-layer perfboard with ground/Vcc planes, and every hole populated with a 2-layer IDC connector (i.e. 4 connections)

Very easy to use, very fast, better electrical performance than wirewrap - but hideously expensive for a hobby.

[KL27x]

Vero wrap is missing the whole wrap part. I find wire wrap is still quite useful for making cables out of pin header. Soldering doesn't provide the strain relief. And with the right stripping technique, there's fairly close to no advantage over regular kynar even when soldering.

99% of the wiring in electronics is very low power signals. 30AWG wire is plenty good for a heck of a lot of repairs and mods and prototyping.

Wire wrapping was another of those products which was 'hyped' without being properly evaluated. Its key problem is that if a connection needs changing then any other wires on top of that connection have to be removed first, and replaced after.

Other than for power and ground, most of the time you only need 2, maybe 3 connections per post/node. It's not often an issue for me, in practice. An extra long header pin can handle 4+ connections, easy, and still leave room for connection to a female header or a probe.

I've never used WW, but it seems as though it would be much easier to check lots of points with a 'scope than it is with a bread-boarded circuit. Once there are a few jumper wires dotted around the place it can hard to find somewhere to clip on a probe. Admittedly, this is without the experience necessary to know the best way to arrange common combos of ICs to optimize interconnecting them.

This is completely true, IME. For anything beyond the very simplest of circuits, I still wire wrap nearly 100% of my connections ON a breadboard. I stick rows of extra long pin header into the board, and any breadboard modules I make have the same long pins which stick up and well as down into the board. 30AWG wires pretty much disappear in comparison to any kind of reuseable jumper wires. When I'm done, I rip off all the wires and throw them away. Each pin gives a place for scope probes, and for female pin headers to stick to. Pin headers are wayyyy better 'en breadboards. The breadboard just gives you a place to stick the headers.   

[Red Squirrel]

I've used wirewrap at my job, I work in a telco CO and sometimes have to punch stuff down on the frame.  Most of the blocks are punch down but we have the oddball block that is wirewrap.  They are neat and there's an odd feeling of satisfaction when you use the wire wrap gun and get a nice connection, but punch down is faster both to perform and to remove. 

If I decided to invest in a wirewrap gun at home I would maybe do it for projects just for fun.  Could probably use PC pin headers. Would actually be faster than soldering small wires when doing point to point on perf boards.

[schmitt trigger]

The real WW-killer has been, in my humble opinion, SMT components. Not only are they significantly are smaller, but one can populate both board sides allowing some extremely efficient and compact designs.

And good double sided PWBAs with solder masks and silkscreens have significantly dropped in price.

[jmelson]

I've never used WW, but it seems as though it would be much easier to check lots of points with a 'scope than it is with a bread-boarded circuit. Once there are a few jumper wires dotted around the place it can hard to find somewhere to clip on a probe. Admittedly, this is without the experience necessary to know the best way to arrange common combos of ICs to optimize interconnecting them.

Yes, IBM even had Tektronix make a thing colloquially called an MST clip.  If you knew the part number, you could order these from Tektronix.  It was a thing that fitted over the tip of a Tek probe, and had a contact that would accept a .025" square pin.  It had a gripping "flag" or handle that made it easy to probe a bunch of points on any wirewrap panel, not just IBM mainframes.  (MST was the name of their 370 computer chips "Monolithic Solid Technology".)

Jon

[thermistor-guy]

Wire-wrap is still a useful technique for some situations. I use it mostly for low-frequency test jigs, test adapters, etc. - situations where I need something right now, and once it works, it will get heavy use, so it must be reliable. At the moment I have a situation where there are no PCBs, just a bunch of connectors fastened to thin steel plate.

Connectors with wrire-wrap terminations can be expensive, or simply unavailable. In the case of D-subminiature connectors, I get parts with the straight PCB pin termination, then wrap to the (round) pins and solder. The soldering needs a deft touch, otherwise the kynar insulation can peel back from the heat. You get the hang of it with practice. As noted by another poster, the wrap gives strain relief.

A few times I have run out of links for 0.1 inch headers on evaluation boards, so I wrapped the (square) header posts instead. Worked fine.

[JonM]

But then Vero Speedwire arrived in the early 80's, which is what I have used ever since.

For some reason I never got on with that.

I did like the IDC version of wirewrap. Imagine a 4-layer perfboard with ground/Vcc planes, and every hole populated with a 2-layer IDC connector (i.e. 4 connections)

Very easy to use, very fast, better electrical performance than wirewrap - but hideously expensive for a hobby.

I would call BICC-Vero Speedwire the "IDC version of wirewrap". I used Speedwire to build some boards with mixed ECL - TTL logic and they worked well. I have a few unused Speedwire EuroCards, the associated sockets and tools waiting for the appropriate application.There was also an ISA bus Speedwire prototype board that had most of the bus interface logic as a printed circuit. I just pulled one out to verify that it was ISA and not a later bus.

I still have an OK Industries battery powered wire wrap tool. I doubt it will ever be used again, but my manual ww tools might be.

[thermistor-guy]

The real WW-killer has been, in my humble opinion, SMT components. Not only are they significantly are smaller, but one can populate both board sides allowing some extremely efficient and compact designs.

And good double sided PWBAs with solder masks and silkscreens have significantly dropped in price.

I would add PC performance and PCB layout s/w to that list.

The layout s/w is now much cheaper, often free, more capable, and available for PCs. PC performance, and especially graphics card performance, have just exploded over the years. That's partly thanks to the high-performance needs of gaming and the size of the gaming market.

I can remember the days of auto-router s/w, with rip-up and retry, just crawling along on an i386 PC, on an overnight run. Gone are the days when you needed an expensive Sun workstation or similar to do any decent layout work.

[joeqsmith]

Some more of my old wire wrapped boards I was using to play around with my own CPU designs.   In the early 80's, I was running a couple of MHz on wire wrap and impressed with my self.  The large board, built in the 90s was using a 100MHz clock.   Fast forward another 20 some years, I am amazed how far we have come.               

https://youtu.be/C8txvmXUIJQ?t=493

Second video showing the larger board in operation. 

https://youtu.be/5OUfx2F43ek

[tggzzz]

But then Vero Speedwire arrived in the early 80's, which is what I have used ever since.

For some reason I never got on with that.

I did like the IDC version of wirewrap. Imagine a 4-layer perfboard with ground/Vcc planes, and every hole populated with a 2-layer IDC connector (i.e. 4 connections)

Very easy to use, very fast, better electrical performance than wirewrap - but hideously expensive for a hobby.

I would call BICC-Vero Speedwire the "IDC version of wirewrap". I used Speedwire to build some boards with mixed ECL - TTL logic and they worked well. I have a few unused Speedwire EuroCards, the associated sockets and tools waiting for the appropriate application.There was also an ISA bus Speedwire prototype board that had most of the bus interface logic as a printed circuit. I just pulled one out to verify that it was ISA and not a later bus.

I still have an OK Industries battery powered wire wrap tool. I doubt it will ever be used again, but my manual ww tools might be.

Same ecological niche, same performance, but I'm referring to a pukka IDC connector on each pin in the matrix.

I occasionally use my manual wirewrap tool, mainly for stripping the 30AWG kynar wirewrap wire.

[@rt]

A lady in a FB group I’m in has made several ISA cards recently, all wire wrap.
I’ve seen some of her older projects, and the only difference is they have seemingly become much neater.
Something I noticed with her very last ISA cards she posted pics of.

I bought the cheap China tool, but still waiting for the right project.
I wire wrapped a core memory address matrix, but that was the only wire wrap in the  project, and still the only WW I’ve ever done.

[JonM]

For the record, wire wrapped and SpeedWire ISA boards from the mid 1980s. SpeedWire was so nice. You could get fully populated (with sockets) boards or just place sockets where you needed them.

Yes, I probably should have used RG-174 or other coax on that SpeedWire version. I don't remember exactly what I used this board for, I think that it was a timing sequencer (9513 was my favorite timer counter chip back in the day, I still have the data book and some chips, 8255 is a GPIO chip).

Both boards were stored in the same storage bin. Some wire wrap pins are smashed together. The Speedwire board appears to be damage free.

[Kleinstein]

With now much better tools to make a PCB and reasonably cheap services to make the boards, there is less need for wire wrapping.

Also more and more parts are SMT only - so they would need extra adapter boards to do wire wrap.
Part size is here less the point - more like availability.

In the early days (1960s) they even used a kind of wire wrapping in small series commercial products. I remember taking apart an old calculator with a back-plane wiring made in a kind of solder-less wire wrapping. Sorry no pictures for that tear-down  . Dave would have loved a 4 banger with nixi tubes and all discrete).
AFAIR the Cray 2 still used that kind of odd technique - fault finding must have been a night mare with a thick layer of all white and blue  (all the same) cables in a chaotic mesh.

[tggzzz]

Yes, I probably should have used RG-174 or other coax on that SpeedWire version.

It looks about 10cm long. Twisted pair with IIRC 1 twist/cm is a good 100ohms transmission line for digital signals. I doubt there would have been significant improvement in signal integrity with that logic's transition times.

[schmitt trigger]

AFAIR the Cray 2 still used that kind of odd technique - fault finding must have been a night mare with a thick layer of all white and blue  (all the same) cables in a chaotic mesh.

I once briefly saw a Control Data's Cyber 70 (I believe) mainframe. You are correct, the amount of wiring was beyond belief.

I had the opportunity to see it because an acquaintance was IT director at this company and invited me to see it during some down time. To improve the clock and bus speeds, Control Data engineers were contracted to refurbish the main processing unit cabinet. The main effort was to re-place and re-route all the wiring.

[Beamin]

And instead of building logic circuits out of individual 74xx chips, you can write some Verilog or VHDL and load it onto a FPGA.

Maybe you can but I can't.  I find the code writing part tedious and frustrating. Why write a line of code in software for say something like debounce when you can add a whole bunch of extra parts.

[duak]

When I was looking for my first job in the 70's, I showed my M6800 micro with 8K of static memory (sixty four 16 pin DIPs) all manually wire wrapped on perfboard.  I made my own bus bars of double sided PCB material for power distribution and made power and ground to the chips a mesh.  I'd like to think it helped.  I gave that computer away but I still have another one based on the 8085 but using newer chips.  Wire wrap was no fun, but it sure beat the alternatives.

We used  wire wrap for prototypes and low volume boards because in the days before computerized PCB layout, because wire wrappers were cheaper than PCB designers.  We could give them a D-sized drawing and usually get a board back in a couple of days, QC'd and everything.  I don't believe I ever found a mistake, something I couldn't always say for myself.  I remember writing a netlist program for backplanes using 8080/Z80 assembler on CP/M.  One computer it ran on had a noisy switching power supply that whistled a neat melody while the line printer banged out the percussion on 11 X 17 fanfold paper.  It beat having to send off and wait for a service center to run the previous Fortran program.

I've got one wire wrap board that's designed for ECL.  It has two level pins and spots for terminators just about everywhere.  I've seen boards using twisted pair for point to point connections.

I understand that Garner Denver developed automated wire wrapping machines in the 60's that were used until the 80's.  I suppose they were programmed by paper tape.

I started using Speedwire in the mid 80's and still have a bunch of new and used cards sitting around somewhere - just in case I decide to build something to use all the DIP chips I saved.  Even then, the boards themselves were generally worth more than the chips they were carrying.

Cheers,

[jmelson]

And instead of building logic circuits out of individual 74xx chips, you can write some Verilog or VHDL and load it onto a FPGA.

Maybe you can but I can't.  I find the code writing part tedious and frustrating. Why write a line of code in software for say something like debounce when you can add a whole bunch of extra parts.

FPGAs are quite different from CPUs.  CPUs do one tiny instructon at a time, everything sequential.  FPGAs and CPLDs do everything in parallel.  VHDL and Verilog may LOOK like computer code, but they are something totally different.  basically, they are a computer-readable schematic.  In fact, at least with Xilinx, you can actually enter a 74xx schematic design and compile it into an FPGA or CPLD device.

The big advantage of these is that you can tinker with experimental changes and view operation in the simulator, without ever having to pull out any wirewrap wires.  If the results are not good, just load the old version back in.

(Yes, you can probably TELL I'm a convert!)

Jon

[RoGeorge]

with Xilinx, you can actually enter a 74xx schematic design and compile it

Enter a schematic was possible in the old Xilinx ISE. The old IDE, Xilinx ISE is still available, but it was frozen, and will not support any current FPGAs, just the old ones.

In their new IDE (new since 2012 or so), Xilinx Vivado, the schematic editor is not present anymore, and AFAIK there were no plans to ever add a schematic editor in Vivado. Did Xilinx added a schematic editor to Vivado?

[NiHaoMike]

In their new IDE (new since 2012 or so), Xilinx Vivado, the schematic editor is not present anymore, and AFAIK there were no plans to ever add a schematic editor in Vivado. Did Xilinx added a schematic editor to Vivado?

Vivado does have a schematic viewer. It even lets you color the individual "wires" as a sort of coloring book for adults. I compiled a miner for my Artix 7 (doesn't make much - about $7 per month) for the fun of it, then an event about a month and a half ago got me to open the design and try to find something to post on Facebook for my friends to see. (They really liked it, BTW.) Then I discovered the "coloring" feature and it went in an unexpected direction... (The design is supposed to help an artist, even if only a little, not become art itself!)

Just one comment from a friend:

Wow! This is so neat and nothing I would’ve ever stumbled upon on my own. Thank you so much for sharing Mike! Your best friend is lucky to have you

[RoGeorge]

The schematic viewer is a completely different thing, and it was always present. We were talking about an editor that enable one to input a circuit description as a graphical schematic diagrams, instead of using text to describe a circuit (like e.g. VHDL or Verilog).

In the old Xilinx ISE it was possible to use graphic symbols, like flip-flops, gates, counters and so on in order to draw the desired circuit. ISE was capable to generate a bitstream starting from schematic drawings. In ISE there are libraries like 7400 series, for those who prefer to design with discrete logic circuits.

Vivado discontinued the schematic input option, and will never add that back.

[woodchips]

Some of my Speedwire AT boards from decades ago.

First two a 68020 version of the Corvus Concept computer.
Three and four a part of a simple sheet metal punch controller I designed, until the WEEE directive decided that 100kg of punch plus 0.5kg of controller was an electronic product so the whole 100kg came under the weight delivered for the payments! Put me out of business.
Five is a combined 68020 with two 56001 DSPs, thought I would continue with image processing when made redundant so made this, never been used!

[RoGeorge]

That's a very clean wire wrap! 

For me, wire wrap sockets were always expensive and hard to find, so even back in the days of the WW glory, I was using wire wrap, but solder the wire instead of wrapping it. I was not even removing the wire insulation, just burned the insulation on the solder pad. This technique of burning the isolation on the pad spares a lot of time with cutting and stripping, yet it was unexpectedly reliable.

I have an entire Z80 computer built with soldered WW, and still works after many decades.



More pics and stories here https://hackaday.io/project/1411-xor-hobby-a-vintage-z80-computer-prototype

Sometimes I still use soldered WW even today, for small prototypes only, but these occasions became less and less frequent.

WW might be still present in very particular cases, but not as a mainstream technique. WW is very expensive, time consuming, not suitable for current IC packaging, it doesn't make sense for SMT, and probably not working at all for high speed buses or high speed differential signals were matching wire lengths and impedances are critical.

I don't think people is still using wire wrap any more.

[daveshah]

Vivado may not have traditional schematic entry, but it certainly has a block design editor that can connect together modules - very useful for dealing with things like AXI buses. I'm sure it has some traditional components too, but they're a bit ugly and it's not really intended for that.

[BBQ]

A few niche products are still manufactured today using wire wrap! Not just old designs, some new too. It seems to last forever done properly.

[jmelson]

with Xilinx, you can actually enter a 74xx schematic design and compile it

Enter a schematic was possible in the old Xilinx ISE. The old IDE, Xilinx ISE is still available, but it was frozen, and will not support any current FPGAs, just the old ones.

In their new IDE (new since 2012 or so), Xilinx Vivado, the schematic editor is not present anymore, and AFAIK there were no plans to ever add a schematic editor in Vivado. Did Xilinx added a schematic editor to Vivado?

Being an old schematic guy, I started with CPLDs and FPGAs using schematics, but quickly realized it was kind of a dead end.  I could put a lot of functionality on one schematic sheet, but I began to see the benefits of an HDL.
So, I'm not sad that schematic -> FPGA has gone away.

Still, on ISE 10 through 14, I'm pretty sure you can still do the schematic thing.  You certainly can do a top-level schematic for the underlying HDL modules (although you don't have to, and I don't.)

Xilinx's ISE 10 through 14 support quite reasonable devices like the 9500XL CPLD family and the Spartan 3A FPGA which are still in production.  These are some of the most affordable parts in the market, too.  If you are not piping serial data streams around at GHz+ speeds, they really do fine, and real mortals can actually solder them to 2-layer boards and have them work.

Jon

Jon

[jmelson]

That's a very clean wire wrap! 

For me, wire wrap sockets were always expensive and hard to find, so even back in the days of the WW glory, I was using wire wrap, but solder the wire instead of wrapping it.

Wow!  A friend of mine built a Data General Nova clone (16-bit mini) with a technique like that.  He had a bunch of wire-wrap-like boards made, but soldered the IC's to pads, and then wired everything together with WWW, but he did strip the ends.  This allowed the boards to pack closer together.

Jon

[RoGeorge]

He had a bunch of wire-wrap-like boards made, but soldered the IC's to pads, and then wired everything together with WWW, but he did strip the ends.

Indeed, with "normal" WW, like the one made in western countries, it was very hard, if not impossible, to make reliable solder joints without stripping. That matte white WW seen in the pic was what we use to have "on the other side" of the Berlin Wall. Not sure where was made that white WW, but I assume somewhere in Eastern Europe, or maybe USSR.

It might have been made in RSR (Socialist Republic of Romania) as well. At that time, any import from western countries was severely limited by the PCR (Romanian Communist Party) if not completely banned, so western WW was like Unobtainium here. To counterbalance the lack of western parts and/or technologies, the Eastern block countries were reinventing the wheel all the time, or simply copy by reverse engineering or industrial espionage. In the particular case of my country, Romania, there was a famous case of a security officer, General Pacepa (the head of Romanian intelligence service), who flee the country during the communist regime. Later, he wrote a book "Red Horizons"
https://www.cia.gov/library/readingroom/docs/CIA-RDP90M00005R001300010017-4.pdf
https://books.google.ro/books/about/Red_Horizons.html?id=zQ7LSucBylAC&redir_esc=y&hl=en
where he described the industrial espionage, including some interesting story about stealing Texas Instruments technology (I think it was about the military 74xx TTL family circuits, but not sure), the plans of a western tank, incredible details about the life of Ceausescu family and other top political figures under the Iron Curtain.

Anyway, back to our eastern European WW, the insulation was very bad when compared to the "normal" western WW. As soon as that matte white insulation was touched with the soldering iron, the insulation burned itself into smoke, yet the solder somehow manage to wet very well the "carbonized" wire, so I used the crappy quality of the insulation as an advantage, in order to avoid stripping and cutting. It was even possible to put a blob of solder in the middle of an unstripped and uncutted wire, which was very handy, especially for memory busses.

[thermistor-guy]

There's just no sense to wire up a bunch of 74xx chips today.

I'm doing just that to make a controller for a bidirectional DC/DC converter. (Actually using 4000 series because that part of the logic runs on 12V.) A FPGA or CPLD would not be able to meet the very low standby/pass through power goal of 1mA or less, plus it would require level translation.

When I was an EE student, one of my classes had a guest lecturer who specialized in industrial process control. He was a big fan of the original 4000 series logic, because of its noise immunity. Running it on 12V or 15 V gave lots of logic level noise margin, plus its slow speed helped it reject fast EMI transients. And it used very little power.

In his applications, the goal was reliability in electrically harsh environments. Most of us were dreaming about how to make computers that ran as fast as possible, but he urged us, in general, to use the slowest logic family that was adequate for the job.

It's interesting to know there are still modern applications for 12V logic.

[tggzzz]

There's just no sense to wire up a bunch of 74xx chips today.

I'm doing just that to make a controller for a bidirectional DC/DC converter. (Actually using 4000 series because that part of the logic runs on 12V.) A FPGA or CPLD would not be able to meet the very low standby/pass through power goal of 1mA or less, plus it would require level translation.

When I was an EE student, one of my classes had a guest lecturer who specialized in industrial process control. He was a big fan of the original 4000 series logic, because of its noise immunity. Running it on 12V or 15 V gave lots of logic level noise margin, plus its slow speed helped it reject fast EMI transients. And it used very little power.

In his applications, the goal was reliability in electrically harsh environments. Most of us were dreaming about how to make computers that ran as fast as possible, but he urged us, in general, to use the slowest logic family that was adequate for the job.

It's interesting to know there are still modern applications for 12V logic.

35 years ago I did a short consultancy study into whether it was worth replacing logic with a micro in a specific application. I concluded that it was not, since the existing logic was extremely reliable and micros would involve all sorts of new hazards specifically including electricity.

That logic was (and probably still is) powered by 2000psi gas.

[NiHaoMike]

It's interesting to know there are still modern applications for 12V logic.

I even managed to condense it down to 3 chips - two doing actual logic and one doing level translation from the 5V microcontroller that generates the PWM signal.

35 years ago I did a short consultancy study into whether it was worth replacing logic with a micro in a specific application. I concluded that it was not, since the existing logic was extremely reliable and micros would involve all sorts of new hazards specifically including electricity.

That logic was (and probably still is) powered by 2000psi gas.

A place where 2000 PSI of high pressure gas is considered less of a hazard than a few volts of electricity must be very strange indeed.

[tggzzz]

It's interesting to know there are still modern applications for 12V logic.

I even managed to condense it down to 3 chips - two doing actual logic and one doing level translation from the 5V microcontroller that generates the PWM signal.

35 years ago I did a short consultancy study into whether it was worth replacing logic with a micro in a specific application. I concluded that it was not, since the existing logic was extremely reliable and micros would involve all sorts of new hazards specifically including electricity.

That logic was (and probably still is) powered by 2000psi gas.

A place where 2000 PSI of high pressure gas is considered less of a hazard than a few volts of electricity must be very strange indeed.

Not really; there are many unmanned offshore oil platform - lots of flammable/explosive gases.

Don't consider how things work, consider how things fail

[jmelson]

A place where 2000 PSI of high pressure gas is considered less of a hazard than a few volts of electricity must be very strange indeed.

Many, many years ago the Moog company (inventors of proportional hydraulic valves) designed an early CNC control.
It used programs punched on wide cloth tape, and the controller was entirely fluidic - compressed air.  The "paper" tape reader looked like something out of a player piano, a plate with many dozens of holes connected to an equal number of little tubes.  The linear encoders on the machine slides were also plates with holes that lined up to holes on a fixed plate in a vernier sort of fashion.  When the holes in the encoder matched the holes in the tape, the axis was on position.  Movement of the axes was done with their proportional valves operating either hydraulic cylinders for short axes or hydraulic motors and leadscrews for the longer ones.

Whew, what a nightmare.  But, the controller was amazingly compact,certainly much smaller than a discrete transistor implementation of the day would have been.

Jon